Challenges associated with autonomous vehicles

Written by: Natasha Kowalskyj, Social Media Summer Work Study
Edited by: Isabella Blandisi-Van Hee, Project Coordinator for Applied Research

By now, these tech blogs have detailed the historical elements, which are necessary for an autonomous vehicle (AV) to function. This blog has also covered the AV implementations happening in Ontario, broken-down how this technology works and what this will mean for the province and Canada as whole, once cities become smarter. The projected benefits have also been explored: environmental, economic, safety, as well as increased accessibility, and what those will provide to the average citizen when cities begin implementing AVs into their future smart cities. These all brought the reader a sense of utopia when learning about AVs being the main mode of transportation and how it will benefit their day-to-day life. However, the reader should be informed about the challenges and necessary changes to everyday life, as they are set to be impactful for many reasons.

The question that first comes to mind with AV implementation is the level of safety the self-driving vehicle will have, not just for the human(s) in the car, but for every pedestrian around these vehicles. Google’s AV is a success story of live road testing. This AV successfully drove itself 800,000+ kilometers on public roads without a single accident (Hanna, 2015). However, not every test of a self-driving vehicle has had the same track record. A year ago, there was an incident where a self-driving Uber [with a human driver present in the car] was being tested on the roads of Tempe, Arizona. Its internal computer did not perceive a human wearing dark clothes crossing its path (with a bicycle) as an avoidable threat and unfortunately, this technical error resulted in the pedestrian being killed by the impact (Somerville, 2018).

Looking at this incident closely, there are a couple of individual faults associated. First and foremost, not to diminish the individual who was struck, but the conditions for which they were crossing the road, can be argued to be dangerous (e.g., dark, no crosswalk); there is the potential that even a human driver may have not had the reaction time to avoid an impact. Once the internal surveillance footage was reviewed, the human driver was accused of being distracted behind the wheel. This point is important because they were required to maintain alertness at all times in case the vehicle required human intervention—it was Level 3 AV or less (see our first post for more information on AV levels). This was the reasoning why the AV did not brake immediately when it sensed the bike—it required human intervention (Somerville, 2018). The accident report concluded that Uber had also disabled the emergency braking system of the AV in question (Somerville, 2018), which is why the car did not respond safely as it should have. It was also concluded that the driver was indeed distracted, only looking up and braking a second after hitting the pedestrian. The Tempe Police Detective, Michael McCormick, required the human driver’s Hulu records to determine when they were on their phone, as the charges of vehicular manslaughter may be laid (Somerville, 2018). This case leads us to consider a discussion surrounding the policing of AVs, determining who is at fault in the event of an accident (especially fatal incidents), and how to insure drivers and AVs alike.

There is also a “grey area” surrounding intoxicated individuals and whether they should receive a punishment for being in the driver’s seat of an AV, the same way one would be punished if they were driving a normal car under the influence. If AVs will be used as a means to reduce drinking and driving (akin to taking taxis or Uber), for example, this question requires a clear answer. Looking at a situation where someone is under the influence, and is sleeping in an AV on their way home, it can be argued that the vehicle is then responsible for maintaining all driving aspects, not the intoxicated passenger. As such, if AVs were to be used in this capacity, we would need to mandate a Level 4 and up AV to avoid incidences like the Arizona case. Then, the question becomes, if a Level 4 or 5 AV is involved in an accident, whose fault is it? Hanna (2015) argues that it is not pertinent to punish an intoxicated individual for being in the driver’s seat of an AV, as it will monetarily cost society, putting more strain on an already overwhelmed justice system while providing little net-benefits. Some States have intensified the testing of AVs and require the human to have a clean driving record—as it relates to driving under the influence—within the last ten years (Hanna, 2015). Ultimately, supporters of autonomous vehicles argue they will make our roads safer by eliminating driving fatalities related to drinking and driving (US Department of Transportation, 2016). If 90% of vehicles on [US] roads were autonomous, the number of accidents would fall from six million a year to 1.3 million—eliminating two-thirds of driving-related deaths (Thompson, 2016).

There are projected challenges when AVs are on the road full-time and how law enforcement will be able to police these vehicles safely. For example, hypothetically, say someone is pulled over for an autonomy inspection and the officer reminds them the roadway they are currently on is certified for Level 2 automation. The police officer would then proceed to ask the driver to show them what level of automation they were driving at before they were pulled over. This poses a potential challenge because if someone was driving at a fully automated level (e.g., 4/5), who is at fault if the vehicle is driving completely autonomously? This becomes a considerable undertaking due to the decisions involving how the level of autonomy for a roadway gets decided. Notably, the discussion of how this will be enforced requires attention as well. The issues of policing, not just the roadways, but also the drivers, as well as deciding who is at fault when there is an accident involving AVs (and potentially distracted drivers), is still a prominent topic in this field. Ending this discussion on a positive note, police believe the use of autonomous vehicles will allow officers to focus their resources/time on more pressing societal matters such as reducing criminal activity (Fink, 2014).

There will also be significant changes to the citywide infrastructure (already in place) and how it will need to be manufactured for AVs to be connected and seamless while driving. The infrastructure necessary to accommodate AVs is expected to be high-tech and impactful at best. However, going through this process of tailoring our cities for the adoption of AVs is expected to save at least 10% on fuel consumption through platooning (as discussed in a previous post), and could also possibly save on public health costs (Duarte & Ratti, 2018). For instance, these costs can be saved through the reduction of road accidents, reducing health problems related to smog/pollution as well as energy consumption (Duarte & Ratti, 2018). While the undertaking of creating new and improved city wide infrastructure for AVs will be disruptive and costly, the implementations of AVs are set to reduce road fatalities by 99% once majority of cars are connected and autonomous (Duarte & Ratti, 2018). These improvements are also set to reduce congestion and improve mobility for underserved populations; all expected to save $60 billion to the US economy alone (Duarte & Ratti, 2018).

Implementations of AVs allow city planners to rethink urban life and the surrounding infrastructure to avoid the pitfalls of poor city planning, and to use this opportunity as a catalyst for what a future city can look and act like. Once drivers are no longer bogged down with the task of commuting, there is the likelihood of more urban sprawl, and with that comes a slew of new issues not previously considered (Duarte & Ratti, 2018). These issues include, but are not limited to: more time travelling to work, leading to increased energy consumption. This may also increase emissions of pollutants (Duarte & Ratti, 2018). Indeed, there is the potential for city centers to become sparsely populated, less active with social interaction, and to have fewer investments into public transportation as it relies on population density to be economically feasible (Duarte & Ratti, 2018). These are only a few of the infrastructure issues expected to become a reality once AVs, and the necessary city transformations, are in fully place.

Like anything in society, the benefits and challenges of embarking on something new has steep learning curves. The projected pros and cons span across both ends of the “disruptive spectrum” and until the time has come to fluidly test these new ways of living connected to vehicles and cities around us, we will not fully know how true or false some of these benefits or challenges are. Until then, cities all around the world will be testing out various levels of AVs (e.g., shuttles, taxis) in order to explore their viability within our day-to-day lives for future generations.


Duarte, F., & Ratti, C. (2018). The impact of autonomous vehicles on cities: A review. Journal of Urban Technology, 25(4), 3-18. doi:10.1080/10630732.2018.1493883

Fink, D. (2014, June 22). Autonomous cars: Driving on auto pilot. Police: The Law Enforcement Magazine. Retrieved from

Hanna, K. (2015). Old laws, new tricks: Drunk driving and autonomous vehicles. Jurimetrics, 55(2), 275-289.

Somerville, H. (2018, June 22). Uber car’s ‘safety’ driver streamed TV show before fatal crash: Police. Retrieved from 

Thompson, C. (2016, December). 8 ways self-driving cars will drastically improve our lives. Retrieved from

US Department of Transportation (2016, September). Federal automated vehicles policy: Accelerating the next revolution in roadway safety [PDF file]. Retrieved from